Background: Fatty accumulation in the rotator cuff is associated with shoulder dysfunction and a risk of failure of rotator cuff repair. The aims of this study were to (1) describe cellular findings in rotator cuff muscles in patients presenting with varying degrees of rotator cuff tendon pathology by examining fat content and myofiber cross-sectional area of rotator cuff muscles and (2) correlate histologic features to magnetic resonance imaging (MRI) grades derived with the Goutallier classification. Methods: Rotator cuff muscle biopsies were performed in a consecutive series of patients undergoing arthroscopic shoulder surgery. Rotator cuffs were graded according to the Goutallier classification and labeled as either partial-thickness or full-thickness. Patients without a rotator cuff tear undergoing arthroscopic surgery served as controls. The biopsy specimens were examined using LipidTOX to visualize lipid accumulation. Laminin was used to quantify myofiber cross-sectional area. Results: Twenty-seven patients with a rotator cuff tear and 12 without a tear (controls) were included. There were 24 males (62%). The mean age was 55 years. Patients in the control cohort were younger (mean, 46 years) than those in the treatment group (mean, 60 years, p < 0.01). Within the treatment group, 12 and 15 patients were recorded as having partial and full-thickness rotator cuff tears, respectively. Lipid accumulation visualized at the cellular level was fairly-to-moderately correlated with the Goutallier classification on MRI (R_s = 0.705, 95% confidence interval [CI] = 0.513, 0.829). Muscle biopsy specimens with a Goutallier grade of 2+ had significantly more lipid accumulation than those with grade-0 (p < 0.01) or grade-1 (p < 0.01) fatty accumulation. Muscle biopsies at the sites of full-thickness tears showed significantly greater lipid accumulation than those associated with either partial (p < 0.01) or no (p < 0.01) tears. Partial-thickness rotator cuff tears had no difference in lipid accumulation in comparison to the control group. Muscle biopsy specimens from full-thickness tears had significantly smaller myofiber cross-sectional area when compared with partial-thickness tears (p = 0.02) and controls (p < 0.01). Conclusions: Cellular lipid accumulation correlates with the MRI Goutallier grade of fatty accumulation, thus verifying the Goutallier classification at the cellular level. Muscle biopsy specimens from partial-thickness tears are more similar to controls than to those from full-thickness tears, whereas full-thickness tears of all sizes showed significantly greater lipid content and smaller myofiber cross-sectional area compared with partial-thickness tears and controls. Clinical Relevance: Our research confirms the utility of using the Goutallier classification to predict rotator cuff muscle quality and shows that tendon attachment, even if partially torn, protects the muscle from fatty accumulation.